During Michaelmas Term, OxTalks will be moving to a new platform (full details are available on the Staff Gateway).
For now, continue using the current page and event submission process (freeze period dates to be advised).
If you have any questions, please contact halo@digital.ox.ac.uk
The AI shutdown problem is, roughly, the problem of getting future advanced (agentic) AI systems to shut down when and only when we want them to. Thornley (2023) proved some theorems showing that an agent satisfying some rather minimal rationality criteria precludes it from being both capable and shutdownable. However, by denying the completeness axiom of utility theory, he manages to characterise a capable and shutdownable agent. A seemingly relevant condition for an agent to be capably goal-directed is that it avoids sequences of actions that foreseeably leave it worse off. My project proposes a choice rule which, as I derive, guarantees the dynamic consistency of agents with incomplete preferences.